The adverse effects of smoking on health have been broadly publicized. Despite this, many adolescents still take up smoking; and many smokers continue to smoke. Several studies of volunteer twin samples have suggested that there may be genetic differences in vulnerability to the smoking habit. A twin-family study is proposed, to explore the role of genetic and social factors, and their covariation and interaction, in (i) the onset of smoking; ii) the persistence of the smoking habit and development of heavy smoking; and (iii) successful smoking cessation vs smoking relapse. Self-report or rating data have already been collected, by mailed questionnaire, on the smoking habits of 2,700 adult twin pairs aged 20-50 from the Virginia Twin Register, and their parents (N=5125), siblings (N=1600) and spouses (N=2960). This proposal seeks support for (a) statistical analysis of these data, by multiple regression analysis and structural equation modelling, and (b) collection and analysis of supplemental data, by mailed questionnaire, from smokers and non-smokers, on (i) personality dimensions which may be predictive of smoking behavior (novelty- seeking, reward-dependence and harm-avoidance), (ii) early history of smoking by self and by cotwin, (iii) exposure to other smokers; and, from smokers and ex-smokers only, on (iv) current consumption. (v) history of smoking cessation and smoking relapse, (vi) subjective reports of craving and withdrawal symptoms, (vii) subjective motives for smoking. The twin family design will permit the joint analysis of genetic and social effects on smoking behavior, including (a) the analysis of genotype x environment interaction (and especially the identification of those social conditions under which genetic vulnerability is exacerbated); (b) the analysis of genotype-environment correlation (arising through (i) environmental influences of the smoking habits of family members (parents, sibs) and (ii) selection by smokers of social environments which permit or encourage smoking - including selection of spouse or friends); (c) the multivariate analysis of the structure of genetic and social influences on smoking behavior (e.g. are there genetically distinct subtypes of smokers?); (d) the analysis of special twin environment effects which could invalidate a tradition twin study.